Gold Market Commentary: Paved with good interventions

A bumper August

A fierce August rally took gold up 13% to end the month at US$4,563/oz. It was the third highest monthly return in a quarter century, narrowly short of the 14% observed in December 2025, and was echoed across currencies (Table 1).

According to our Gold Return Attribution Model (GRAM), momentum factors – led by widespread ETF buying – were a major contributor, followed by a weaker US dollar via opportunity cost FX. A rise in implied gold volatility, representing not risk but strong call buying, also helped drive prices higher (Chart 1).

ETF flows were strong in August, with contributions across the board. Europe narrowly pipped the US, recording inflows of US$7.9bn (54t), followed by North America with US$7.8bn (53t); Asian gold ETFs recorded US$2bn (13t) of inflows.

COMEX net managed money positions increased by 97t (US$13bn) in August. The other reportable category, likely reflecting Commodity Trading Advisor (CTA) activity, rose by 115t (US$17bn). 

Paved with good interventions

• The recently announced US Treasury buybacks are officially about liquidity.¹ But some are seeing them more as financial repression: an attempt to stop yields rising²
• Real assets like gold will likely continue to benefit, reflecting investor concerns over growing deficits and debt, until a credible plan to deal with them is formulated.

In August, we opined on the US Treasury's surprise buyback announcement and the potential narrowing path to yield curve control. The buyback increase kicks off on 9 September.³ As investor and media interest hasn't waned, we wanted to revisit in a little more detail.

Intervention chatter did not start with the buyback announcement but arguably a fortnight earlier, with the rather odd assistance the Treasury extended to Japan during its own intervention in early August.⁴ Maybe it was the knee-jerk manner, or that it landed straight after hawkish Fed minutes. Or perhaps it was that it coincided with US debt hitting the big 40, normally a milestone worth celebrating.⁵

Regardless, the story has been hogging headlines ever since.  And a strong public riposte to the Treasury’s intervention from investor Stanley Druckenmiller likely echoed a wider frustration with a US administration unwilling to apply more than ‘sticking plasters’ to a much more serious problem.⁶

What form intervention takes, or even who conducts it, is perhaps less relevant than how the market views it. Although the US Treasury has considerable fire power, the Fed’s is unlimited…should they decide to get involved. Nominal yields would almost certainly be capped. But where would the pressure go instead? If the market takes it in its stride, then perhaps nowhere. If it interprets intervention as desperate, the release valve would likely be falling real yields, an expanding term premium and a lower US dollar – or a crowding out of private sector demand for these assets. Real success would necessitate a credible deficit-reduction plan.

There is no useful precedent for this, so our analysis is only suggestive. 

We look at weekly data from 2000 and the moves in nominal yields, real yields, term premium, dollar index and gold. 

We filter the data on two themes: a credible intervention and a confidence-eroding one, recording gold's excess return (less its full sample average) in the event week, the following week and two-to-four weeks on. 

Under each theme, nominal yields are filtered to be ‘fixed’ within a range as we assume that intervention succeeds on its own terms. What differs is the composition beneath: whether the move comes out of the term premium, or out of real rates while inflation compensation builds, or out of the US dollar on capital outflows. Chart 2 shows our assumptions and the filters we have used. 

Theme one: Credible intervention

This theme sees financing concerns ease: real yields sit flat to slightly lower and the term premium compresses. With deficits still at war-time levels, inflation worries don’t disappear, hence real yields are a bit softer than nominal yields. Everything else is held inside a standard deviation move, which is how a market taking intervention in its stride might look.

Theme two: Confidence-eroding intervention

This theme is not quite the mirror image. Nominal yields are contained as before, but the other three are filtered on moves by up to three standard deviations. Real yields fall as inflation fears build, the term premium no longer compresses, and the dollar weakens as capital looks elsewhere. It is the combination, rather than any one leg of it, that marks this out as financial repression rather than a garden-variety risk-off week. There is as mentioned, the possibility term premia – which are model derived, real yields and even the dollar are distorted. As such the willingness or lack thereof of holding assets would not be visible in prices but in flows.

Results

Chart 3 shows how gold fared during these thematically filtered weeks:

• Pinned nominal yields are quite unusual when the other components move: of 1,443 weeks since 2000, only 30 fit the credible theme and 20 the confidence-eroding one
• Unsurprisingly, gold fares far better under a confidence-eroding intervention than a credible one. Falling real yields and a weaker dollar are typically bread and butter to gold’s short-run returns. But this does underline gold’s value as a hedge against precisely that risk
• A successful intervention is not necessarily bad for gold. Mechanical success doesn’t solve the underlying problem – and through moral hazard could make it worse. A compressing term premium on its own is insufficient to dent gold’s excess return
• However, one warning shot across the bow is that the two weakest returns in the credible theme sample – May 2014 and July 2015 – coincided with a US deficit that had fallen to around 2.5% of GDP from 4% eighteen months earlier. Two observations prove little, but capping yields alongside genuine fiscal restraint represents a near-term risk for gold, even if remote. 

In summary

In the end it boils down to whether the market is convinced. If it isn't, gold probably benefits. If it is, then one of the narrative supports for gold’s strong multi-year run is temporarily lost. In our view, spending and tax commitments make that a challenging proposition.

While the preceding analysis is focused on the US, rising yields – particularly given high debt burdens – are a global concern (Chart 4). Like a game of whack-a-mole, a temporary solution here could worsen the problem elsewhere as investor flows look for a home. Next on the agenda is a more likely Fed hike in September (Chart 5). Solid economic data and billowing inflation fears had suggested it in June, before softer data allayed those fears. But now expectations have see-sawed on a pervasive continuation of the US-Iran conflict, bolstered by Fed Chair Warsh’s hawkish comments and the recent Jackson Hole Symposium. 

What a hike achieves isn’t clear and on paper wouldn’t be great for gold. We laid bare some of the possible dynamics in an earlier commentary, noting that it is not the yield move itself but what it embeds that matters. A hike could restore policy credibility and flatten the curve. That’s the sort of scenario that could confound the simple rule that hikes are always bad for gold. Let’s see.